Stress hyperglycemia and associated glycemic variability are common in critical care and have been found to contribute to infection, slow wound healing, and short and long-term mortality. Early clinical trials utilizing intensive insulin therapy to maintain Tight Glycemic Control (TGC) in surgical patients, resulted in a 34% decrease in in-hospital mortality and comparable decreases in morbidity. Consequently, many hospitals adopted aggressive inpatient glycemic targets for their Intensive Care Unit (ICU) patients. However, in subsequent studies the benefits of TGC were less apparent. Many factors contribute to the variability of reported TGC outcomes, including (i) the insulin therapy protocols themselves, which offer surprisingly divergent insulin dose recommendations under similar circumstances, (ii) the etiology of extreme BG variability in diverse patient populations, and (iii) the accuracy of the methods used to assess BG concentration and inform the insulin therapies. The main obstacle to developing optimized insulin therapy protocols is that clinical trials are expensive, time consuming, and cannot compare glycemic outcomes for different protocols in identical patients. To overcome these problems we have proposed and partially validated an innovative Bioengineering Insulin Therapy Design Methodology (BITDM) based on (i) characterization of the population-specific variability in patients' insulin sensitivity parameters in a well validated simulation model of glucose metabolism; (ii) creation of a computer-based ICU BG Simulator centered around a virtual subject population that replicates the responses of the real patient population to different insulin therapies; and (iii) simulator-based in silico design and optimization of TGC algorithms. The central hypothesis of this proposal is that BITDM can provide an accurate means of evaluating, comparing, and optimizing insulin protocols that safely improves TGC in specific patient populations, in this instance cardiovascular surgery (CVS) patients. Accordingly, the following specific aims are proposed: Aim 1. Apply BITDM to the distinct physiology of postoperative CVS patients. Aim 2. Establish the efficacy of BITDM-based insulin protocol optimization within U.Va.'s cardiothoracic ICU. Once established, BITDM will allow for different glycemic management strategies to be tested, compared and optimized within a common set of in silico patients at a fraction of the cost and time it will take to perform a clinical trial. If successful, this study will allow us to start addressing our long-term goal - developing a simulator-based bioengineering methodology for optimization of insulin treatment that safely achieves TGC and allows for improved outcomes.